Make scanned pages searchable

A scanned page is a picture of a page. You cannot search it or select its text. From Omitly 5.0, a redacted scan can carry an invisible text layer, so you can search the saved PDF and select text on it.

What it does

If Omitly read a page with OCR during your session, it can add the layer after redaction. It works in this order:

1. Omitly redacts the page as usual.
2. It reads the redacted page image again.
3. It writes the recognised text invisibly behind that image.

The text comes from the redacted pages, never from the original document. Anything under a redaction is removed before the reading starts, and Omitly checks that no recognised word lies inside a redaction.

Pages that already contain real text are not affected.

Turn it on or off

Go to Settings → Redaction → Make scanned pages searchable. It is on by default from 5.0. Switch it off if you want the saved PDF to stay a plain picture of each page.

When Omitly adds it

Omitly adds the layer when you use:

• Apply & verify, for regions you drew and regions found by OCR.
• Auto-detect personal info.
• Full-page redaction.
• Replace and name lists.
• AI-proposed redactions that you accept.

It does not add the layer in batch runs. It does not add it when Encrypt output (password protection) is on.

It is checked before it is saved

Before saving, Omitly reads the output again. If the invisible text does not match what it can read back from the redacted pages, it refuses to save and tells you why. Nothing is written.

What it can get wrong

The layer is recognised text. It may be incomplete or wrong. Treat it as a convenience for searching, not as a copy of what the page says. In 5.0 it recognises English only.

What you see after redacting

The result shows Searchable: N words on pages …. It also lists:

• Pages where no text was recognised.
• Pages that were not read, with a reason: could not be rendered, rotation not valid, cancelled, or timed out.

It never shows you the recognised words.

What the certificate says

The certificate and the .audit.html report state:

• Which pages carry the layer.
• That the recogniser’s name, version, language and confidence floor are the producer’s testimony. Omitly does not verify them.
• That recognised text is not the original and may be incomplete or wrong.
• That no recognised word lies inside a redaction. Omitly checks this by recomputing word positions from the delivered file. It states this only when the check passed.
• Which pages were not examined.

For what the certificate does and does not prove overall, see What the redaction certificate proves.

Privacy

Everything happens on your device. The document never leaves it. See Where does your PDF go?.